Genotypic Diversity of Complement Component C4 Does Not Predict Kidney Transplant Outcome

Abstract

Gene copy number of complement component C4, which varies among individuals, may determine the intrinsic strength of the classical complement pathway. Presuming a major role of complement as an effector in transplant rejection, we hypothesized that C4 genetic diversity may partially explain the variation in allograft outcomes. This retrospective study included 1969 deceased-donor kidney transplants randomly selected from the Collaborative Transplant Study DNA bank. We determined recipient and donor gene copy number of total C4, C4 isotypes (C4A and C4B), and C4 gene length variants (C4L and C4S) by quantitative real-time PCR analysis. Groups defined according to recipient C4 gene copy number (low, intermediate, and high) had similar 10-year allograft survival. Genotypic groups showed comparable rates of graft dysfunction, treatment for rejection, immunological graft loss, hospitalization for infection, malignant disease, and death. Similarly, separate analyses of C4A, C4B, C4L, and C4S; combined evaluation of donor and recipient C4 genotype; or analysis of recipients with higher risk for rejection did not reveal considerable outcome effects. In conclusion, we did not demonstrate that C4 gene copy number associates with transplant outcome, and we found no evidence that the resulting variation in the strength of classical complement activation influences susceptibility to rejection.

The importance of the complement system in infective and inflammatory disease has been redefined over the past decade, and its role has been extended to many new situations including allograft rejection.^1,2 Much of the new knowledge derives from studying knock-out mice and genetically deficient humans, and little is known about the effect of subtle variations in complement function. However, this is now important because targeted manipulation of the complement system is now feasible and could be used as therapy, for example in organ transplantation.^3,4

Complement has a powerful influence on allograft rejection in murine models, which also emphasize the predominant influence of locally synthesized complement.⁵ In this context, the third complement component (C3) was demonstrated to play a key role, but it is not clear which pathway activates it in rejecting allografts.^5–7 The fourth complement component (C4) is central to activation of the classical complement pathway,⁸ and in clinical transplantation, endothelial deposition of its cleavage product C4d correlates closely with antibody-mediated renal allograft rejection.^9,10 This unique association between C4d deposition and graft injury suggests a pathogenetic role in rejection.^2,11

The C4 gene is located on chromosome 6 in the MHC as a member of the RP-C4-CYP21-TNX (serine/threonine kinase RP, complement component C4, steroid 21-hydroxylase, and tenascin X [RCCX]) module. The C4 locus has a complex genetic organization caused by the combination of: (1) substantial copy number variation (CNV) of RCCX modules and hence interindividual differences in the number of C4 gene copies; (2) gene diversification resulting in two isotypes (C4A and C4B), either of which can be missing; and (3) dichotomous size variation leading to short (C4S) and long (C4L) genes.¹² Importantly, the number of copies inherited correlates closely with plasma C4 concentrations and functional complement activity^13,14 and may contribute to susceptibility to infection and autoimmune disease.^15–18 Consequently, genetic analysis in kidney transplant donor recipient pairs provides an unusual human model system to study the influence of C4 in immune-mediated injury that uniquely is able to distinguish between local and systemic origin.

Here we examined whether C4 gene CNV (and hence C4 concentrations and activity) affects kidney allograft survival. The study used nearly 4000 randomly selected samples from the Collaborative Transplant Study (CTS) biobank that contains DNA from a large number of donor recipient pairs archived together with long-term outcome data. Thus we were able to undertake a study with sufficient power to ascertain definitively whether or not C4 gene CNV either in the donor or the recipient influences long-term organ allograft outcomes.

RESULTS

C4 Genotyping Results

In this cohort study, 1969 kidney transplant recipients and their corresponding donors (n = 1946) were subjected to C4 genotyping. Applying quantitative real-time PCR analysis, we determined the individual gene copy number for the two major C4 isotypes C4A and C4B and, in parallel, their two length variants C4S and C4L. Total C4 gene dosages were calculated as the sum of C4A and C4B genes. High accuracy of our approach using quantitative PCR is supported by a high level of concordance (99.2%) with the sum of the C4S and C4L genes. Donor and recipient cohorts had comparable C4 gene copy number distributions, which were strikingly similar to an earlier published cohort of European ancestry.¹⁴ The mean copy number per diploid genome was 3.88 ± 0.73 (means ± SD) (recipients) and 3.83 ± 0.72 (donors), respectively. The most common number of C4 copies was four for both recipients and donors (n = 1166 [59.2%] and n = 1150 [59.1%] of tested subjects, respectively), followed by three (n = 458 [23.3%] and n = 499 [25.6%]), five (n = 262 [13.3%] and n = 212 [10.9%]), two (n = 55 [2.8%] and n = 55 [2.8%]), six (n = 25 [1.3%] and n = 28 [1.4%]), seven (n = 2 [0.1%] and n = 1 [0.1%]), or eight (n = 1 [0.1%] and n = 1 [0.1%]) C4 gene copies, respectively. The results for C4 isotypes and length variants are listed in Table 1.

Table 1.

Recipient and donor gene copy number of C4 variants

Gene Copy Number	Number of Recipients (%)	Number of Donors (%)
C4A genes
0	31 (1.6)	18 (0.9)
1	310 (15.7)	330 (17.0)
2	1125 (57.1)	1154 (59.3)
3	432 (21.9)	391 (20.1)
4	62 (3.1)	49 (2.5)
5	8 (0.4)	3 (0.2)
6	1 (0.1)	0 (0.0)
7	0 (0.0)	1 (0.1)
C4B genes
0	61 (3.1)	51 (2.6)
1	498 (25.3)	519 (26.7)
2	1260 (64.0)	1237 (63.6)
3	138 (7.0)	129 (6.6)
4	12 (0.6)	9 (0.5)
5	0 (0.0)	1 (0.1)
C4L genes
0	23 (1.2)	12 (0.6)
1	138 (7.0)	135 (6.9)
2	420 (21.3)	449 (23.1)
3	702 (35.7)	776 (39.9)
4	607 (30.8)	515 (26.5)
5	73 (3.7)	53 (2.7)
6	6 (0.3)	5 (0.3)
7	0 (0.0)	1 (0.1)
C4S genes
0	715 (36.3)	702 (36.1)
1	869 (44.1)	839 (43.1)
2	318 (16.2)	354 (18.2)
3	56 (2.8)	45 (2.3)
4	11 (0.6)	6 (0.3)
5	0 (0.0)	0 (0.0)

Baseline characteristics of recipients with low (<4), intermediate (4), or high (>4) C4 gene copy number are shown in Table 2. The common European 8.1 ancestral haplotype (HLA-A1, B8, DR3) carries a single C4B and no C4A gene.^19,20 Consequently, it was highly prevalent among recipients and donors with low C4 gene dosage (39.4 and 26.8%, respectively), which resulted in their having better HLA-matched grafts (Table 2).

Table 2.

Baseline characteristics, according to recipient total C4 gene copy number

Characteristic	C4 Gene Copy Number			Pa
	<4 (n = 513)	4 (n = 1166)	>4 (n = 290)
Geographic origin, number (%)				0.045
Europe	442 (86)	1049 (90)	263 (91)
North America	71 (14)	117 (10)	27 (9)
Transplant year (means ± SD)	1995.9 ± 5.3	1996.2 ± 5.2	1996.5 ± 5.2	0.32
Original renal disease, number (%)				0.42
glomerulonephritis	190 (37)	455 (39)	123 (42)
polycystic kidney disease	64 (12)	139 (12)	37 (13)
diabetes mellitus	59 (12)	94 (8)	24 (8)
Recipient gender, number (%)				0.34
female	215 (42)	451 (39)	108 (37)
male	298 (58)	715 (61)	182 (63)
Recipient age, years (means ± SD)	46.4 ± 14.1	46.8 ± 13.6	47.7 ± 13.9	0.31
Retransplantation, number (%)	72 (14)	152 (13)	35 (12)	0.72
Donor gender, number (%)				0.44
female	203 (40)	473 (41)	105 (36)
male	310 (60)	692 (59)	183 (64)
Donor age, years (means ± D)	39.2 ± 17.1	40.9 ± 16.6	41.3 ± 15.8	0.18
Cold-ischemia time, hours (means ± SD)	20.7 ± 8.2	20.2 ± 8.4	20.0 ± 7.8	0.39
HLA mismatch (A, B, DR), number (%)				<0.001
0 to 1	94 (18)	116 (10)	23 (8)
2 to 4	361 (70)	882 (76)	216 (74)
5 to 6	58 (11)	168 (14)	51 (18)
Recipients typed HLA-A1, -B8, and -DR3, number (%)b	202 (39)	26 (2)	1 (0)	<0.001
Panel-reactive antibody, number (%)				0.89
0 to 5%	312 (65)	706 (65)	179 (67)
6 to 50%	87 (18)	214 (20)	51 (19)
>50%	79 (17)	169 (15)	38 (14)
Initial immunosuppression, number (%)				0.7
CsA + AZA + steroids	204 (40)	475 (41)	120 (41)
CsA + MPA + steroids	99 (19)	223 (19)	46 (16)
CsA + steroids	90 (18)	189 (16)	44 (15)
Tac + MPA + steroids	57 (11)	113 (10)	35 (12)
otherc	63 (12)	166 (14)	45 (16)
Induction antibody therapy, number (%)				0.83
IL-2 receptor antibody	48 (9)	101 (9)	21 (7)
depleting antibody	133 (26)	292 (25)	71 (24)

AZA, azathioprine; CsA, cyclosporin A; MPA, mycophenolic acid; Tac, tacrolimus.

^aP values are given for comparisons among all three groups.

^bHLA antigens corresponding to the 8.1 ancestral haplotype.

^cThe rate of each other combination is less than 3%.

Recipient C4 Gene Diversity and Renal Allograft Outcome

The recipients were subdivided into three categories according to C4 gene CNV (<4 copies, 4 copies, and >4 copies) for clinical analysis. Patient groups showed virtually identical graft survival, death-censored graft survival, or patient survival over a 10-year follow-up period (Figure 1). Also multivariate Cox regression analysis did not reveal any differences (Table 3).

Figure 1.

Recipient C4 gene copy number does not associate with kidney allograft survival rates. Rates of graft survival (A), death-censored graft survival (B), and patient survival (C) are shown according to recipient C4 gene copy number group.

Table 3.

Recipient C4 gene copy number and clinical outcome - multivariate Cox regression analysis

Recipient Gene Copy Number	Number of Transplants	Graft Survival		Death-censored Graft Survival		Patient Survival
		Hazard Ratio (95% CI)	P	Hazard Ratio (95% CI)	P	Hazard Ratio (95% CI)	P
C4
<4	513	1.00 (0.85 to 1.18)	0.99	1.06 (0.86 to 1.31)	0.59	0.94 (0.75 to 1.19)	0.62
4	1166	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
>4	290	0.95 (0.77 to 1.17)	0.62	0.92 (0.70 to 1.20)	0.54	1.01 (0.77 to 1.33)	0.95
C4A
<2	341	1.00 (0.82 to 1.21)	0.99	0.97 (0.76 to 1.25)	0.84	1.08 (0.84 to 1.40)	0.56
2	1125	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
>2	503	1.04 (0.88 to 1.23)	0.68	1.00 (0.81 to 1.23)	0.99	1.09 (0.87 to 1.37)	0.45
C4B
<2	559	1.04 (0.89 to 1.22)	0.64	1.10 (0.90 to 1.34)	0.36	0.97 (0.78 to 1.21)	0.81
2	1260	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
>2	150	0.94 (0.72 to 1.24)	0.68	0.93 (0.65 to 1.33)	0.70	1.05 (0.73 to 1.50)	0.80
C4L
<3	581	0.98 (0.82 to 1.17)	0.81	1.04 (0.83 to 1.30)	0.76	0.98 (0.77 to 1.25)	0.89
3	702	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
>3	686	0.94 (0.79 to 1.11)	0.48	1.02 (0.82 to 1.26)	0.88	0.97 (0.77 to 1.22)	0.82
C4S
0	715	1.03 (0.88 to 1.20)	0.74	1.14 (0.94 to 1.40)	0.18	0.97 (0.78 to 1.20)	0.75
1	869	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
>1	385	1.04 (0.86 to 1.26)	0.65	1.10 (0.86 to 1.40)	0.47	0.99 (0.76 to 1.28)	0.91

CI, confidence interval.

We then analyzed the effect of C4 gene CNV on the relative distribution of five major categories of graft loss documented in the CTS database, namely: graft failure for unclear reason, failure from rejection, nonimmunological (including technical) failure, recurrence of original disease, and death with functioning graft. Graft loss from rejection and death with function were the most common causes of graft failure (29 and 37% of all documented losses, respectively), whereas graft loss from disease recurrence was rare (10 cases; recurrent glomerulonephritis: n = 4). As shown in Figure 2, there was no significant difference between the three genotypic groups in the causes of early (first year, P = 0.88) or late graft loss (years 2 to 10, P = 0.33). Specifically, recipients with low C4 gene copy number did not have fewer immunological failures or more graft losses from recurrent disease, as we had anticipated (Figure 2).

Figure 2.

Recipient C4 gene CNV does not influence the relative distribution of five major categories of graft loss: graft failure for unclear reason (Unclear); failure from immunological rejection (Immunol); nonimmunological failure (Non-Immunol; technical failure included); recurrence of original disease (Rec Dis); and death with functioning graft (Death). The results obtained for the three recipient C4 genotypic groups are shown for the first year (A) and years 2 to 10 (B). Comparisons of groups for individual causes of graft loss did not reveal statistical significance.

The 1-year serum creatinine values were similar in all patient groups (P = 0.87) (Figure 3), and there were no significant differences in the number treated for rejection (P = 0.18). Similarly, recipient C4 gene copy number had no influence on rates for hospitalization for infection (P = 0.31) during the first year or deaths from cardiovascular death (P = 0.12) and malignant disease (P = 0.64) up to 10 years after transplantation. Finally, recipient C4A, C4B, C4L, or C4S gene copy number had no effect on overall graft survival, death-censored graft survival, or patient survival (Cox regression analysis; see Table 3). The distribution of causes for graft loss was similar for all genotypic groups, and they had comparable graft function, rates of anti-rejection treatment, hospitalization for infection, malignancies, and deaths from cardiovascular disease (data not shown).

Figure 3.

Recipient C4 gene copy number does not associate with renal allograft function at 1 year.

Effect of Donor C4 Gene Copy Number on Transplant Outcome

To examine the possible contribution of donor C4 gene expression, we categorized transplants according to both recipient and donor C4 gene copy number (recipient and donor <4, recipient ≥4 and donor <4, recipient <4 and donor ≥4, and recipient and donor ≥4 genes). Graft and patient survival rates were similar among these four groups (Figure 4). Causes of early or late graft loss were similarly distributed among the four genotypic groups, and there was no considerable difference in terms of graft loss from rejection within the first year (P = 0.98) and between years 2 and 10 (P = 0.21), respectively. Similarly, there were no differences in allograft function at 1 year (P = 0.44) or rejection treatment during the first year (P = 0.25). To assess the effect of donor genotypic diversity of C4 variants, transplants were categorized according to recipient and donor C4A, C4B, C4L, or C4S dosage. Univariate analysis of C4 variants showed no difference in graft or patient survival (data not shown), and the only difference revealed by multivariate Cox regression analysis was a lower hazard ratio for death-censored graft loss (0.73 [confidence interval, 0.56 to 0.94; P = 0.016] in recipients with ≥1 C4S genes receiving a C4S null kidney compared with those with a ≥1 C4S gene donor (Table 4).

Figure 4.

Recipient and donor C4 gene copy number do not associate with survival rates. The rates of overall graft survival (A), death-censored graft survival (B), and patient survival (C) are shown according to recipient-donor genotypic group.

Table 4.

Recipient and donor C4 gene copy number and clinical outcome - multivariate Cox regression analysis

Gene Copy Number			Number of Transplants	Graft Survival		Death-censored Graft Survival		Patient Survival
	Recipient	Donor		Hazard Ratio (95% CI)	P	Hazard Ratio (95% CI)	P	Hazard Ratio (95% CI)	P
C4	<4	<4	226	1.04 (0.83 to 1.31)	0.71	1.21 (0.91 to 1.60)	0.19	0.95 (0.70 to 1.29)	0.76
	<4	≥4	287	1.01 (0.82 to 1.24)	0.94	1.09 (0.83 to 1.42)	0.55	0.87 (0.65 to 1.16)	0.35
	≥4	<4	334	1.05 (0.86 to 1.28)	0.63	1.26 (0.99 to 1.60)	0.055	0.85 (0.64 to 1.12)	0.24
	≥4	≥4	1122	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
C4A	<2	<2	152	1.18 (0.91 to 1.52)	0.22	1.16 (0.83 to 1.61)	0.39	1.25 (0.90 to 1.76)	0.19
	<2	≥2	189	0.88 (0.68 to 1.12)	0.30	0.86 (0.62 to 1.20)	0.39	0.96 (0.69 to 1.33)	0.79
	≥2	<2	201	1.10 (0.87 to 1.38)	0.44	1.13 (0.84 to 1.51)	0.41	1.25 (0.92 to 1.70)	0.15
	≥2	≥2	1427	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
C4B	<2	<2	186	1.06 (0.82 to 1.35)	0.67	1.28 (0.95 to 1.72)	0.11	0.72 (0.49 to 1.06)	0.094
	<2	≥2	373	1.04 (0.86 to 1.25)	0.71	1.11 (0.87 to 1.41)	0.40	1.01 (0.79 to 1.30)	0.92
	≥2	<2	391	0.99 (0.83 to 1.20)	0.95	1.20 (0.95 to 1.51)	0.13	0.82 (0.64 to 1.06)	0.13
	≥2	≥2	1019	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
C4L	<3	<3	282	1.13 (0.93 to 1.39)	0.23	1.16 (0.89 to 1.50)	0.27	1.10 (0.84 to 1.45)	0.50
	<3	≥3	299	0.90 (0.73 to 1.12)	0.35	0.94 (0.71 to 1.23)	0.65	0.88 (0.66 to 1.18)	0.39
	≥3	<3	321	1.02 (0.84 to 1.25)	0.82	1.07 (0.83 to 1.38)	0.60	0.96 (0.74 to 1.27)	0.80
	≥3	≥3	1067	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)
C4S	<1	<1	332	0.97 (0.79 to 1.18)	0.74	1.05 (0.82 to 1.34)	0.70	0.90 (0.67 to 1.20)	0.47
	<1	≥1	383	0.95 (0.78 to 1.15)	0.60	0.98 (0.77 to 1.25)	0.89	0.97 (0.75 to 1.26)	0.84
	≥1	<1	380	0.83 (0.68 to 1.01)	0.059	0.73 (0.56 to 0.94)	0.016	0.90 (0.70 to 1.17)	0.45
	≥1	≥1	874	1.00 (reference group)		1.00 (reference group)		1.00 (reference group)

CI, confidence interval.

Clinical Effect of C4 Gene Copy Number in Presensitized Patients

Next we performed a subanalysis of 301 patients with >5% panel reactivity to ascertain whether C4 gene dosage influenced graft survival in recipients with increased risk of antibody-mediated and presumably complement-mediated transplant injury. Even in this cohort, recipient C4 gene copy number had no effect on graft survival (Figure 5). Similarly, analyses of C4A, C4B, C4L, or C4S dosage or combined evaluation of recipient and donor C4 genotype did not reveal significant differences (data not shown).

Figure 5.

C4 gene copy number in presensitized recipients does not associate with renal allograft survival.

DISCUSSION

The results from this large study provide no evidence that recipient and/or donor C4 gene CNV influences long-term renal transplant outcomes and argue strongly against C4 genotyping for risk stratification in organ transplantation. This demonstrates that the variation in the intrinsic strength of the classical complement pathway conferred by C4 gene CNV is insufficient to influence susceptibility to allograft rejection. Our negative result contrasts with other immune-mediated inflammatory settings, such as susceptibility to autoimmune disease and the response to infection, both of which have been associated with C4 gene CNV.^15–18

The large sample size is a major strength of our study and gives it sufficient statistical power to justify our conclusions. However, it did require the use of archived samples and retrospective database analyses with all their inherent limitations. Inevitably, some relevant outcome variables were not recorded in the CTS database, and in particular there have been many developments in the biopsy-based classification of rejection type and grade since the CTS DNA archive and database were established. Furthermore, the majority of our recipients were transplanted before the introduction of C4d staining and the introduction of accurate biopsy-based diagnostic criteria for acute and chronic antibody-mediated rejection into the Banff classification.²¹ This prevented us from correlating C4 gene CNV with T cell-mediated and antibody-mediated acute and/or chronic rejection as currently diagnosed. Despite these caveats about the definitions of the diagnoses recorded in the CTS database, we found no significant effect of C4 CNV on “treatment for rejection” or “graft failure due to immunological rejection,” whether early or late, even in sensitized recipients at particular risk of antibody-mediated rejection. These data provide powerful indirect evidence against C4 gene CNV having a major effect on (antibody-mediated) rejection, even though the database-derived end points are imprecise and most probably aggregate different types of rejection. This conclusion is consistent with the observation that C4 gene CNV had no effect discernable on long-term graft survival up to 10 years after transplantation, either in the whole cohort or in sensitized patients. Nor was there a shift in distribution of the frequencies of the five major categories of graft loss in the CTS database and specifically no statistical support for the expectation of less rejection and more recurrent disease in the low C4 gene dosage groups. The lack of influence on long-term survival strongly argues against a clinically relevant influence of C4 gene CNV on the severity of antibody-mediated rejection in the light of recent reports suggesting that chronic antibody-mediated injury is one of the most common features in late graft dysfunction and a major determinant of long-term survival.^22–24

The use of transplantation provides a unique opportunity to distinguish between the consequences of C4 gene CNV on systemic and local production of C4, which is important in light of studies demonstrating local C4 production in human kidney tubular cells.^25,26 We found little support for an effect of donor C4 genotype on outcome, even when analysis was confined to recipients at the extremes of C4 CNV. The only significant difference found on multivariate analysis was a small decrease in death-censored graft failure for C4S null kidneys transplanted into recipients with ≥1 C4S genes. The C4S variant is associated with more effective gene transcription than the C4L variant,^13,14 raising the possibility that a donor C4S null genotype impairs local C4 synthesis, thus limiting complement-mediated injury. However, this is unlikely because C4S null kidneys transplanted into C4S null recipients were not similarly protected. Interpreting our single significant P value, it is important to note that even if the null hypothesis (no effect) is true, P values of ≤0.05 can be expected to arise by a chance about 1 in 20 times. Regarding the considerable number of statistical tests in the multivariate analysis of C4 variants, and thus a high chance of getting one or more spurious significant differences, the significant P value in our analysis is likely consistent with a false-positive result.

Our transplant cohort included a considerable number of individuals with hetero- or homozygous deficiency of C4 variants, but none with complete C4 gene deficiency, as expected because of its rarity.^27,28 Accordingly, our results do not exclude a mechanistically important role for C4 in the transplant rejection. However, they do exclude a clinically discernable effect on allograft outcome, and by implication rejection, of the substantial variation in the efficiency of C4 activation caused by C4 gene CNV.^13,14 It is possible that even the background of a very low gene copy number in recipients and/or donors is above the threshold required for efficient classical pathway complement activation. Nonetheless, in light of recent experimental and clinical studies addressing the role of complement in transplantation, the negative results of our study may be of considerable practical interest.

In human transplantation, there is currently no direct proof for a causal relationship between local complement activation and graft injury. For example, although a marker of antibody-triggered complement activation, capillary C4d deposition, is not invariably associated with rejection, and some have even suggested that its occurrence in stable allografts could indicate protection from injury.^29,30 Conversely, antibody-mediated tissue injury does not depend uniquely on complement activation and can be associated with complement-independent alterations of the integrity of graft endothelium.^31–33 Experiments performed in complement-deficient rodents have generated some controversy regarding the contribution of classical complement activation to alloimmune injury. In a mouse kidney transplant model, recipient or donor C4 deficiency was reported to have no effect on the course of allograft rejection.⁶ Moreover, C1q deficiency in mice turned out to even enhance antibody-mediated rejection of heart allografts.⁷ Finally, passive transfer of donor antigen-specific antibody into heart-transplanted mice was reported to induce chronic injury also in the absence of full complement activation and local C4d deposition, respectively.³³ These experimental data are consistent with an anecdotal report of three kidney allograft recipients, who experienced rejection episodes despite complete C4 deficiency.³⁴

One can speculate that a lack or impairment of classical complement activation could be compensated for by other complement pathways, such as C3 activated through the alternative pathway. Indeed, in rejecting C1q- or C4-deficient rodents, local deposition of the C3 cleavage product C3d was noted,^6,7 and there is experimental evidence for a role of C3 as a critical checkpoint linking complement to components of innate and adaptive immunity.⁵ A recent human study of the influence of C3F, C3S allelic variants in renal allografts reported superior survival of C3F/F or C3F/S kidneys transplanted into C3S/S recipients.³⁵ This was interpreted as showing a relevant contribution of the C3 component to chronic graft injury, which is considerably affected by the functional differences conferred by the C3F/S polymorphism. However, the results were not replicated in a subsequent better-powered study evaluating 1147 donor and recipient pairs selected from the CTS study.³⁶ In this study, C3 polymorphism did not exert any effect on graft outcomes. This study significantly adds to this negative report and strongly supports the previous conclusion that graft rejection and injury is a multifactorial process that is unaffected by the variation in complement brought about by common polymorphisms and CNV of the C3 and C4 genes. Both studies using the CTS were sufficiently powered to provide the definitive conclusion that genotyping of the two major complement components C3 and C4 has no prognostic value in the context of human kidney transplantation.

Previous relatively small studies have reported that complete or partial C4 deficiencies increase susceptibility to both infectious disease^15,16 and myocardial infarctions and strokes.³⁷ Our much larger study failed to replicate either association for C4 gene CNV in the 10 years of follow-up after transplantation, at least as reflected by the recorded data on cardiovascular death and hospitalization for infection. One cannot exclude the possibility that a relatively small effect on C4 gene CNV found in the general population is overwhelmed by the much more powerful additional risks present in the renal transplant population.

In conclusion, this study did not reveal a significant effect of C4 gene CNV on kidney allograft outcome and provides indirect evidence that intrinsic strength of classical complement within the range influenced by C4 gene CNV does not affect the susceptibility of renal allograft recipients to acute or chronic transplant rejection. These results argue against a diagnostic benefit from recipient and donor C4 genotyping for risk stratification in transplantation.

CONCISE METHODS

Study Design and Patients

The DNA used in this retrospective study has been collected prospectively after written informed consent at transplantation units participating in the CTS study. Collection of biologic material and prospective recording of clinical data were approved by the local ethics committees of participating centers, and this study was performed after approval by the ethics committee of the University of Heidelberg (application 083/2005Ä). To maintain anonymity, genotyping (performed at the Medical University of Vienna) and statistical evaluation of data (performed at the University of Heidelberg) were carried out in a blinded fashion.

Recipient and donor pairs were eligible for inclusion if they fulfilled the following criteria: (1) deceased donor kidney transplantation, (2) transplantation between 1988 and 2006 in Europe or North America; (3) no multiorgan transplantation; (4) Caucasian recipient and donor; (5) no missing values for original disease, recipient, and donor age, graft number, or HLA typing; and (6) known type of immunosuppression (intention to treat). In a first step, 2070 recipient-donor pairs were randomly chosen from the DNA bank of the CTS study. For 3946 of the selected samples, sufficient DNA was available, resulting in a total of 1969 transplants, for which complete genotyping was possible for both recipient and donor. These transplants, which had been performed at 55 transplant centers in 13 countries, were finally included. There was no difference in graft survival between selected patients and 25,074 recipients who fulfilled the initial criteria but were not genotyped (P = 0.96), indicating that sample selection was indeed random and representative. With approximately 2000 recipients and a significance level of 0.05, this study has a test power of 80% to detect a relative risk of ≤0.8 or ≥1.25.

C4 Genotyping

Genomic DNA was extracted from peripheral blood (recipients) and lymph node or spleen samples (donors). C4 gene dosage was assessed by quantitative real-time TaqMan® PCR analysis refined by minor groove binder technology (Applied Biosystems, Rotkreuz, Switzerland). Real-time PCR analysis was performed in 384-well optical plates on a 7900HT fast real-time PCR system (Applied Biosystems, Rotkreuz, Switzerland). Primers and probes specific for C4A, C4B, C4L, and C4S (common C4A and C4B forward primer “C4totalfwd”: 5′-GCA GGA GAC ATC TAA CTG GCT TCT-3′; common C4A and C4B reverse primer “C4totalrvs”: 5′-CCG CAC CTG CAT GCT CCT-3′; probe “C4A”: FAM (carboxyfluorescein)-ACC CCT GTC CAG TGT TAG; probe “C4B”: FAM-ACC TCT CTC CAG TGA TAC; common C4L and C4S forward primer “C4Fin95”: 5′-TTG CTC GTT CTG CTC ATT CCT T-3′; reverse primer “C4L-3LTR-R”: 5′-GTT GAG GCT GGT CCC CAA CA-3′; reverse primer “C4Sin9R-2”: 5′-GGC GCA GGC TGC TGT ATT-3′; and common C4L and C4S probe “C4in95”: FAM-CTC CTC CAG TGG ACA TG) were designed according to previously published protocols.^38,39 Each of the four distinct PCR batches contained TaqMan® Universal PCR Master Mix, No AmpErase® UNG (uracil-DNA glycosylase) (ABI catalog number 4326614), VIC (2′-chloro 7′-phenyl-1,4-dichloro-6-carboxyfluorescein)-conjugated TaqMan® RNase P control reagents (ABI catalog number 4316844), 250 nM of the respective FAM-conjugated TaqMan® probes (C4A, C4B, C4L, or C4S), the particular primers (C4A and C4B: 300 nM; C4L and C4S: 750 nM), and autoclaved MilliQ pure water. Appropriately prediluted genomic DNA (threshold cycle [C_T] values for RNase P between 24 and 30) was added before start. Thermal cycler conditions were adjusted as follows: initial denaturation step for 10 minutes at 95°C; 40 cycles including denaturation for 15 seconds at 95 °C; and annealing/extension for 1 minute for 60°C (exception: C4S at 59°C). The data were analyzed using SDS 2.3 software (Applied Biosystems, Rotkreuz, Switzerland).

Fifty genomic DNA samples from consanguineous subjects were purchased from the International Histocompatibility Working Group and used as positive control DNA (http://www.ihwg.org/cellbank/dna/refpan_hla_consang_table.html [May 18^th, 2010]). The gene dosages of these individuals have earlier been characterized by two independent methods, Southern blot and real-time PCR.³⁹ The C_T value of RNase P and each complement parameter (C4X = C4A, C4B, C4L, or C4S) was converted into a raw gene dosage (which does not take into account the different PCR efficiencies of RNase P-VIC and the respective C4-FAM assay) by the formula nRAW_C4X = 2^{(C_TRP)−(C_TC4X)+1}. Raw gene dosages of positive controls selected from the consanguineous reference panel were plotted versus the actual gene dosages, and the resulting calibration curve served for determination of the actual copy number of unknown samples of this particular run.

Statistical Methods

Fisher's exact tests were used to compare groups of categorical data. Kaplan Meier analysis was applied for calculation of graft and patient survival. The mantel Cox Log-rank test was used for comparison of survival between groups. For multivariate analysis, Cox regression analysis was performed using the following confounders: year of transplantation, geographical region (continent), first or repeat transplant, recipient and donor sex and age, cold-ischemia time, number of HLA mismatches (A, B, DR), preformed panel-reactive antibodies, original disease leading to end stage renal failure, and immunosuppressive therapy (intention to treat). Hazard ratios are presented with 95% confidence intervals. A two-sided P value of <0.05 was considered statistically significant. Statistical calculations were performed using IBM SPSS Statistics (version 18.0, SPSS Inc., Chicago, IL).

DISCLOSURES

None.